The Nonlinear Characteristics Of Tropical Climatic Variability And The Nonlinear Relationship Between Ocean And Atmosphere

Based on the monthly mean of OLR, divergence field at 925hPa and 200hPa level, wind field at 925hPa and 200hPa level and sea level pressure of NCEP/NCAR reanalysis datasets, nonlinear spatial correlation patterns of convective activity , divergence field and southern oscillation are researched by applying nonlinear principal component analysis method. The results show that:(1) Average climatic maps show that equatorial belt of converage is quite different along to the longitude at the same season in location, and its displacement is also different along to season. At north hemisphere the belt is northward during summer while it is southward during winter. (2) The EOFAI model includes annual and interannual variability, the first mode of EOFAI model show the spatial patterns of seasonal differences of two hemispheres, which is consistant with the results of average climatic. The EOFI model only includes interannual anlmalies, the first mode of EOFI is very similar to the second mode of EOFAI. (3) Taking EOFI analysis to the anomolysis of OLR, field of divergence at 925hPa and sea level pressure, the results show that EOF describes only the standing oscillation associated with ENSO variability, but it cannot explain lower dimensional nonlinear characteristics. (4)The nonlinear principal component analysis method are applyed to the three spatial patterns, the results of spatial distributions are quite different at intensity and spatial patterns while PC is set maximum and minmum. The strongest negative OLR anomalies , the strongest negative lower troposphere anlmalies and negative sea level pressure anomalies during an average El Nino event lie closer to the eastern boundary of the tropical zone than do the coldest anomalies during an average La Nina event, and the intensity during an average El Nino event are stronger than during an average La Nina event. While different values of α₁(t_n)are chosen to provide a representative sample of spatial structures associated with the NLPCA approximation, it is found that NLPCA mode describes the evolutions of average ENSO events, and asymmetry characterizes are prestended during the whole cycle of ENSO. (5) By using composite analysis method to the three spatial fields, the composite distributions show the antisymmetric spatial during El Nino and La Nina, and the symmetric and antisymmetric components of the composite El Nino and La Nina also indicate their linear and nonlinear responses to ENSO.